51.- Thermal Coupling Tutorial

This is my first tutorial fully written on english, then I beg your pardon for any sintax error or mistake I made have — thanks!.

In the following link you have a FEMAP & NX NASTRAN Tutorial explained step-by-step dealing with Thermal Coupling of Steady-State Heat Transfer Finite Element Analysis from CHIP to PCB (Printed Circuit Board) with GLUE Surface-to-Surface Thermal Contact and Orthotropic Material Properties:


I have created in FEMAP the following picture that explains the problem very well:


The loads & boundary conditions are the following:

  • A heat load Q= 5 Watt applied to the top surface of the chip.
  • A temperature constraint of 25°C applied to the edge of the PCB opposite to the edge closest to the chip.
  • A convection coefficient applied on the top and side faces of the chip of 24 W/m2ºC.
  • A convection constraint to the top face of the PCB with a heat transfer coefficient = 19 W/m2ºC, being 25ºC the ambient temperature of the fluid around the PCB.
  • A thermal coupling between the CHIP & PCB, assuming the chip is connected to the PCB using a Ball Grid Array (BGA) solder connection with a total area that is half of that of the chip.
  • Assume as well a Pb-Sn weld of thermal conductivity Ksolder= 80 W/mºC and a gap of 1 mm. We can then calculate the total conductance as follows:

Heat Transfer Coefficient =
ksolder * Asolder / (Lgap * Achip) =
80*0.5/0.001*1 = 40e3 W/m2ºC

For simplicity, I suggest to use always GLUETYPE=1 and PENTYP=2, which allows you to directly specify the conductance via PENN, having the following units:

(Thermal Conductivity*Length)/Area

Thermal Conductivity in SI has units of W/mºC, thus PENN has units of W/m2ºC, then we can use directly the value of Heat Transfer Coefficient for the Thermal Coupling.

The PENN value is independant of the GLUE surface area, the user does not need to know the glue surface area by advance, this value is computed by the glue algorithm and is a function of the mesh size & shape, glue refinement, etc..

So, setting Glue Type = 1..Spring (GLUETYPE=1), also Penalty Factor Units = 2…Force/(Length x Area) (PENTYP=2) and Normal Factor = 0.04W/mm2ºC (PENN), the definition of GLUE Thermal Coupling parameters will be correct as shows the following image:


The following picture shows the final steady-state results of temperature distribution in the Chip & PCB components where you can see the effects of the orthotropic material in the temperature distribution: Heat is better conducted in the Y orthotropic material direction. Temperatures vary from approximately 25ºC to 60ºC.


The following picture shows the temperature gradient in the chip, with the values on the color bar limited to those found in the Chip only by means of using Groups:


I hope the above tutorial to be helpful & useful.
Best regards,

Datos de Contacto de IBERISA (Spain)

4 thoughts on “51.- Thermal Coupling Tutorial

  1. Hola Blas
    Muy interesante y útil, el próximo calculo térmico lo voy hacer siguiendo esta guía.
    Pero me ha salido una duda. ¿Por qué utilizas un factor de convección diferente para el chip (24W/m2C) y para la PCB (19W/m2C)?
    Si no recuerdo mal el valor para la convección natural está entre 5-25W/m2C


  2. Estimado David,
    La teoría matemática de la convección de calor es muy complicada, no existe ninguna ecuación sencilla como ejemplo para conducción. Ello es debido al hecho de que el calor ganado o perdido por una superficie a cierta temperatura en contacto con un fluido a otra temperatura distinta depende de muchas circunstancias, a saber:
    1.- De que la superficie sea plana o curva, pulida o rugosa.
    2.- De que sea horizontal (mirando hacia arriba o hacia abajo) o vertical.
    3.- De que el fluido en contacto con la superficie sea un líquido o un gas.
    4.- De la densidad, viscosidad, calor específico y conductividad térmica del fluido.
    5.- De que la velocidad del fluido sea lo suficientemente pequeña para producir un régimen laminar o lo bastante grande para originar un régimen turbulento.
    6.- De si tiene lugar evaporación, condensación o formación de películas, etc…

    El coeficiente de convección “h” como puedes ver es altamente variable, y complicado de calcular, y no existe mucha información al respecto. A veces es necesario realizar un análisis de fluidos CFD para calcular precisamente la distribución del factor de película!!. En resumen, que el tema es complicado, y tú mismo estás dando un rango de variación entre 5-25 W/m2ºC para convección natural que me parece razonable. En mi caso he estimado un factor de película “medio” mayor para el CHIP que para el PCB en base a los factores anteriores: temperatura del chip mayor, temperatura ambiente mayor, incluye caras verticales y horizontales, etc..



  3. If you tried to assume that you had no solder and simply placed a chip on top of a PCB so that it was in simply contact, how would you account for this? Sorry, I don’t speak much Spanish and would probably make the problem worse if I tried. Thanks.


  4. Dear Gabriel,
    Please take a look to my website at the address: http://www.iberisa.com/soporte/femap/termico/Heat_Transfer_from_Chip_to_PCB.htm, there you have explained in detail the capabilities of the NX NASTRAN GLUE surface-to-surface contact.

    The glue control parameters on the BGPARM bulk entry can help you adjust the glue algorithm. The DEFAULT values of GLUE contact defines a PERFECT GLUE contact condition, ie, in Heat Transfer Analysis a perfect conductivity between coupled entities, but if you require to TUNE the thermal coupling then you will need to “play” with the different glue control parameters of BGPARM entry (see NX NASTRAN QUICK REFERENCE GUIDE, Remark#2 of BGPARM entry).
    Best regards,


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s